Inspiration

With the dawn of the internet and the Web, much of the world's information has become freely available to most people in the world. Now, the question of becoming educated is not so much a problem of access to high-quality information or even high-quality educators, but rather a question of what you will spend your time and energy learning. With an abundance of information, new challenges arise like deciding what information you will "pay" attention to as well as dealing with the growing complexity of this mass of information. Sifter aims to help with both of these new problems that arise due to information overload.

What it does

Sifter is a web extension that ingests the information you feed it through simple text selection.

What it will do

With the text selection, and the surrounding context, Sifter will identify noun phrases and the relationships between noun phrases to create a detailed graph structure of the information you read.

When zooming out, the most high-level topics should be most visible, and when zooming in, more detail should populate the graph. In this way, the user can explore their knowledge easily in a non-linear as well as from a top-down or bottom-up perspective.

Technologies being used

Sifter uses web technologies: TypeScript/JavaScript, CSS3, and HTML5.

Challenges

The main challenge is abstracting from the natural language text, high-level descriptions providing a top-level perspective of the user's knowledge.

Another big challenge will be to create a dynamic graph visualization to display high-level nodes as well as allow you to zoom in to view the more detailed nodes in the knowledge graph.

What's next for Sifter Web Annotator

Reasoning over the knowledge graph for a number of use cases: to incorporate new knowledge from the web into a central knowledge graph, to be able to provide you insight in your personal knowledge graph, providing you improved content recommendations, and helping you actively review information you have learned/viewed by generating questions related to your knowledge base.

Share this project:

Updates